Transcript presentation

Assessing euro area residential
property prices through the lens of
key fundamentals*
L. Gattini
European Central Bank
December 2011
* This presentation represents the views of the presenter which do not necessarily
correspond with those of the ECB.
Outline
1. Background
2. Data and modelling approach
3. Results
4. Other approaches for the euro area & additional results
for countries
5. Conclusions
2
Background
Two approaches to the analysis of house price developments:
•
Analysis of the boom and bust cycles – Kennedy and
Andersen (1994), Alessi and Detken (2009), Borgy et al.
(2009) – costly and not costly cycles
•
Understanding the contribution of the fundamental
components of housing market to price developments –
Tsatsaronis et al. (2009), Goodhart and Hofmann (2008)
Mixed set of potential explanatory variables
•
•
Real economic variables
“Credit view”: credit, leverage, overall bank balance sheet
•
Interest rate and risk taking channel
3
Data and modelling approach
“Forecasting and assessing euro area house prices through
the lens of key fundamentals” – Gattini & Hiebert, ECB WP
1249
• An empirical framework for the analysis of house prices in the
euro area using a vector error-correction model (VECM)
• Real house prices related to selected housing demand and
supply fundamentals
• We employ a structural decomposition – its is more suitable for
policy analysis and interpretation of results in light of
equilibrium/dis-equilibrium
4
Data and modelling approach
• Estimated on quarterly data – 1970/2010
• Sources: ECB, OECD, Eurostat
• Data for some countries are interpolated (e.g. DE)
• 4 variables: Real Disposable Income, Real Mixed Interest
Rate, Real Private Residential Investment, Real House Prices
• Vector Error Correction model – parsimonious specification
• 5 lags – Akaike Schawarz and FPE criteria
• 1 Cointegrating Relation – Johansen Method – unrestricted
cointegration rank tests (trace and maximum eigenvalue)
5
Data and modelling approach
6
Data and modelling approach – literature using ECM
approach
Abelson et al. (2005)
Hunt and Badia
(2005)
Meen (2002)
Meen (2002)
Jacombsen and Naug
(2005)
OECD Economic
Survey (2004)
OECD Economic
Survey (2004)
Min
Model
Max
Income
Interest
Rates
Housing
Supply
Country
Period
Freq.
1.7
-5.4
-3.6
Australia
75-03
Q
1.91
-6.0
-
UK
72-04
Q
2.7
2.5
-1.3
-3.5
-7.9
-1.9
US
UK
81-98
69-96
Q
Q
1.7
-3.2
-1.7
Norway
90-04
Q
1.94
-7.14
-0.52
Netherlands
79-02
Q
3.6
-4.5
-7.3
Spain
89-03
Q
1.7
3.07
3.6
-7.14
-6.87
-1.3
-7.9
-2.21
-0.52
Euro Area
70-08
Q
7
Modelling approach – Structural decomposition
- (S)VECM system more suitable for policy analysis - structural
decomposition is useful to analyse the responsiveness of the
system – k*(k − 1)/2=6 restrictions imposed
- Common trend approach – King, Plosser, Stock and Watson
(1991), Iacoviello (2002)
- We distinguish between structural shocks with permanent and
transitory effects
Permanent shocks - baseline identification imposes zero
restrictions on the first and second columns of the D(1) elements
8
Modelling approach – Structural decomposition
Housing market technology shock
- Technological shocks to the construction industry - rarely
observed
- Motivated by changes in the regulatory framework (e.g. building
regulations and/or the modification of various zoning laws)
- This could cause changes in housing production virtually
indistinguishable from housing building technology - Matsuyama
(1999)
9
Modelling approach – Structural decomposition
The impact on real interest rates would be ambiguous
• The euro area - relatively closed economy
- A substitution effect between categories of investment should
nullify possible discrepancies in terms of returns between different
categories of investment in the long-run
- Sectoral specific technology shocks can have an impact in the
short-run on interest rates – zero in the long-run given
counterbalancing effects
10
Modelling approach – Structural decomposition
Economy wide technological shock
- Expected to exert some impact on all the variables in the
system
Financing cost shock
The outcome of features that permanently alter interest rate risk
premia, such as financial innovation or –specific to the case of
euro area– convergence in the run up to European Monetary
Union.
11
Modelling approach – Structural decomposition
Transitory shocks - A two way short-run interaction between real
interest rate and real income has been excluded via imposing two
zero restrictions - standard lags in the monetary policy
transmission mechanism
Housing demand shock - The temporary shift in preferences
toward housing assets
- Rationalized in the context of literature on a time-varying housing
risk premia (seeWeeken, 2004)
- A temporary shift from non-residential demand to residential
demand
12
Results – Forecast as a testing tool
13
Results – Impulse Response – housing demand shock
14
Results – Impulse Response – housing supply shock
15
Results – Historical decomposition
Real house price
Fin. Cost Shock
Economy Wide Technological Shock
Housing Market Technology Shock
Housing Demand Shock
Real house price
0.20
Real housing investment
Fin. Cost Shock
Economy Wide Technological Shock
Housing Market Technology Shock
Housing Demand Shock
Real housing investment
0.20
0.15
0.15
0.10
0.10
0.05
0.05
0.00
0.00
-0.05
-0.05
-0.10
-0.10
-0.15
-0.15
2009Q3
2008Q1
2006Q3
2005Q1
2003Q3
2002Q1
2000Q3
1999Q1
2009Q3
2008Q1
2006Q3
2005Q1
2003Q3
2002Q1
2000Q3
1999Q1
16
Results – Transitory-permanent component – B-N type
Real House Price
0.4
FIN. COST Component
Real house price
INCOME Component
INVEST Component
HP Component
Real house price (RH axis)
Perm. comp. of real house price (RH axis)
Fin Cost Shock
INCOME Shock
INVEST Shock
HP Shock
Transitory Component
0.3
0.2
0.1
5.0
5.0
4.4
4.9
3.8
4.8
3.2
0
4.7
2.6
4.6
2.0
-0.1
4.5
1.4
-0.2
4.4
0.8
4.3
0.2
-0.3
2010Q3
2008Q1
2005Q3
2003Q1
2000Q3
1998Q1
1995Q3
1993Q1
1990Q3
1988Q1
1985Q3
1983Q1
1980Q3
1978Q1
1975Q3
1973Q1
2010Q3
2008Q1
2005Q3
2003Q1
2000Q3
1998Q1
1995Q3
1993Q1
1990Q3
1988Q1
1985Q3
1983Q1
1980Q3
1978Q1
1975Q3
1973Q1
-0.4
17
4.2
Results – Transitory-permanent component – B-N type
FIN. COST Component
Real house price
INCOME Component
INVEST Component
HP Component
Real house price (RH axis)
Perm. comp. of real house price (RH axis)
CREDIT Shock
Real House Price
INCOME Shock
INVEST Shock
HP Shock
Transitory Component
0.15
0.1
0.05
5.0
5.0
4.4
4.9
0
3.8
-0.05
4.8
3.2
4.7
2.6
-0.1
4.6
2.0
-0.15
4.5
1.4
-0.2
4.4
0.8
-0.25
4.3
2010Q2
2008Q1
2005Q4
2003Q3
2001Q2
1999Q1
1996Q4
1994Q3
1992Q2
-0.4
1990Q1
2010Q2
2008Q1
2005Q4
2003Q3
2001Q2
1999Q1
1996Q4
1994Q3
1992Q2
1990Q1
0.2
18
4.2
Other approaches for the euro area
• Crude affordability in the euro area – measured by the ratio of
per capita GDP to the house price index – computed relative to
long-term trends
• Residual of a simple error-correction framework with real house
prices regressed on real GDP per capita, population and the real
interest rate
• House price-to-rent ratio computed relative to its long-run
average – a simplified static dividend discount model or asset
pricing approach
• The evolution of the house price-to-rent ratio computed relative
to the real long term interest rate - return on a housing investment
should be equal to the returns on alternative investment
opportunities
19
Other approaches - euro area - MB
20
Other approaches - euro area countries - FSR
21
Conclusions
• Proposed a structural methodology capturing fundamental
demand and supply factors
• The structural model suggests that euro area housing has been
overvalued in recent years – 2006/2007 by 10% -, implying a
period of stagnation, which is already started in 2009, to bring
housing valuation back in line with its modelled fundamentals
• During the last house price boom much of the increase appears
to reflect a permanent component with an increasing importance
of real disposable income per capita
• A transitory component has also contributed – particularly
since 2006. In particular, housing preference and income shocks
were a key driver in explaining house price dynamics over this
period
22
Thank you for your attention !!!